decreasing proliferative capacity theory

Aubrey de Grey ag24 at
Fri Feb 9 10:16:06 EST 2001

Sydney Shall wrote:

> However, in Science last week there were two papers from the MRC labs at
> UCL in London England, which reported that although fibroblast cells did
> senesce, glial and oligodendocytes did not!!!!!
> A remarkable result.
> In last week's Nature Thea Tlsty has a paper that says that cells
> eventually do senesce.  The field is once again confused with data!!!

Sydney, I would be very interested if you (as undoubtedly the resident
expert on cell senescence on this newsgroup) could expand on why you
consider these papers to be as remarkable as you suggest.  The Science
articles were mentioned on when they came out, and
I said this:

> These are very interesting studies, but not nearly so amazing as it
> sounds here.  The cells in question, like many rodent cell types,
> were expressing telomerase, so the Hayflick limit as it's generally
> thought about doesn't apply.
> It's therefore worth describing why the result is interesting at all.
> Mouse cells, as I guess most readers of this group know, undergo the
> same process of replicative senescence as human cells -- and, indeed,
> at a much smaller number of divisions -- despite generally having much
> longer telomeres than human cells.  The prevailing theory for why they
> do so is that cell cycle arrest is brought on not only by critically
> short telomeres but by other DNA damage -- most probably double-strand
> breaks -- and that in the normal environment of a cell culture, which
> has a much higher oxygen concentration than what exists inside the
> body, the frequency of such damage is enough to cause the cell to stop
> trying to fix it as it happens and instead to initiate a fall-back
> plan, cell cycle arrest, that (almost always) stops the cell from
> accumulating mutations which could turn it into cancer.  DNA repair in
> human cells is much better than in mouse cells, so they can keep going
> even in that high oxygen until their telomeres begin to matter.  Mouse
> cells, on the other hand, can't do this.  I know of unpublished work
> that confirms a very strong prediction of this theory: if you grow
> mouse and human cells in low oxygen, you should get a rise in both in
> the number of divisions before replicative senescence, but that rise
> should be much bigger in mouse cells than in human.  From memory, the
> rise was about 30% in human cells and 200% in mice.  So this new rat
> result is really just an even better confirmation of that same theory.

Similarly, the Nature article you mention, concerning human mammary
epithelial cells, seems to me to establish that these cells are similar
to mouse fibroblasts -- that is, they have a telomere-independent early
senescence mechanism which is rather easy to escape from.  This is a
nice counterpart to the similarity between mouse germ cells and human
fibroblasts, where there is no barrier to growth arrest until the
telomere-dependent one (demonstrated in telomerase knockout mice and
ES cells).

What all this adds up to, for me, is a fairly economical generalisation
that covers all types of cultured cells from either mice or humans, and
which is not challenged by any of the recent articles mentioned above:

1. Telomere maintenance (in most cells in the culture) is required for
   truly indefinite proliferation.

2. So is adequate chromosomal maintenance (in most cells in the culture)
   -- "adequate" meaning fast enough to keep up with the rate of damage.

3. Telomere maintenance is absent in human cell culture but frequently
   present in mouse cell cultures (either at isolation or soon after)
   unless they are made robustly telomerase-negative (by knockout).

4. Adequate chromosomal maintenance is present in human fibroblast cell
   culture but absent in mouse cell culture unless the medium and/or the
   oxygen tension is made unusually permissive.  It's also absent in
   human mammary epithelial cells.

5. Whichever of telomere shortening or chromosomal damage hits the buffers
   first, something close to classical cell senescence occurs (unless the
   machinery for it has already been broken, e.g. by p53 knockout).

6. If it was chromosomal damage that triggered cell senescence and there
   is still a way to go before telomeres matter, escape into a second
   rapid-growth phase is relatively easy.  Thereafter, when the telomeres
   come to matter, the cell senescence response has already been eluded
   so something messier (crisis) happens instead.

7. If, on the other hand, telomeres are the first problem, senescence is
   still available, and it happens, and escape only occurs if telomerase
   -- or possibly ALT -- is reactivated (which in human culture typically
   requires prior, forced escape from senescence using, e.g., SV40).

Is this as consistent with the data as I'm claiming?  One prediction of
it is that with better culture conditions, human mammary epithelial cells
would not exhibit the first growth plateau at all -- that they only do
so because they're somewhat less good than fibroblasts at coping with
chromosomal damage.  Is this (or its negation) known?

Aubrey de Grey

More information about the Ageing mailing list